Original Paper Pathobiology 2010;77:96–105 DOI: 10.1159/000278291
Received: August 17, 2009 Accepted after revision: November 16, 2009
Expression and Alteration of p16 in Diffuse Large B Cell Lymphoma Ee-Soo Lee a Lian-Hua Kim c Wan Ariffin Abdullah b Suat-Cheng Peh a Departments of a Pathology and b Paediatrics, Faculty of Medicine, University of Malaya, and c Department of Mechatronic and BioMedical Engineering, Faculty of Engineering and Science, University of Tunku Abdul Rahman, Kuala Lumpur, Malaysia
Key Words p16 ⴢ Diffuse large B cell lymphoma ⴢ Immunohistochemistry ⴢ Fluorescence in situ hybridization
DLBCLs show different patterns of P16 expression. High levels of P16 may mitigate tumour cell proliferation. Gains of p16 do not necessarily increase P16 protein expression. Copyright © 2010 S. Karger AG, Basel
Abstract Objectives: This study aimed to examine (1) the expression of P16 protein relative to sites of presentation, immunophenotypic subgroups and proliferative indices of tumour cells, and (2) the relationship between p16 gene alterations and P16 protein overexpression in 70 cases of diffuse large B cell lymphoma (DLBCL). Methods: Expression of P16, CD10, BCL6, MUM-1 and proliferation marker (Ki-67) was demonstrated by immunohistochemistry. Fluorescence in situ hybridization (FISH) was employed to detect p16 alterations. Results: P16 overexpression was shown in 45.7% (32/70) of the DLBCL cases, and was significantly correlated with CD10 (p = 0.022) and germinal centre B-cell-like (GCB) phenotype (p = 0.022). High expression of P16 was inversely associated with high proliferative activity (Ki-67 index greater than 75%) (p = 0.020). Of the 47 cases that yielded interpretable FISH results, 57.4% (27/47) showed deletions of p16 and 27.7% (13/47) showed gains of p16. P16 overexpression and p16 deletions were mutually exclusive (p = 0.019). There was no correlation between P16 overexpression and p16 gains (p = 0.621). Conclusions: The GCB and non-GCB subgroups of
© 2010 S. Karger AG, Basel 1015–2008/10/0772–0096$26.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/pat
Introduction
Progression of cell cycle is tightly controlled at multiple checkpoints to ensure normal cell proliferation. Passage through each checkpoint is regulated by a series of cyclins, cyclin-dependent kinases (CDK) and CDK inhibitors [1]. p16 (also known as CDKN2A, MTS1 or INK4A) is a tumour suppressor gene that resides in the 9p21 chromosomal region [1, 2]. Expression of p16 has first been discovered as a 16-kDa protein that binds exclusively to CDK4 and CDK6 [3]. In the proliferative state, CDK4 and CDK6 are activated by cyclin D to phosphorylate the PRb. Phosphorylated PRb releases E2F, a transcription factor responsible for cell cycle transition from G1 to S phase. Binding of P16 to CDK4 and CDK6 may inhibit phosphorylation of PRb and induce cell growth arrest at the G1 checkpoint [1, 2]. There is controversy concerning the expression of P16 in carcinogenesis. Loss of P16 as a result of genetic inactivation has been a common event in tumour cell lines and primary tumours [2, 4]. However, varying results have Ee-Soo Lee Department of Pathology, Faculty of Medicine University of Malaya 50603 Kuala Lumpur (Malaysia) Tel. +60 3 7967 7539, Fax +60 3 7955 6845, E-Mail esoolee @ gmail.com
Materials and Methods CD10 +
–
GCB (n = 23)
BCL-6 +
– Non-GCB (n = 22)
MUM-1 + Non-GCB (n = 21)
– GCB (n = 4)
Tissue Samples A total of 70 formalin-fixed, paraffin-embedded specimens from DLBCL cases (34 nodal and 36 extranodal) obtained between the years 1996 and 2006 were retrieved from the archives of a single institution. All cases were reconfirmed as DLBCL based on the criteria of the WHO classification of haematopoietic and lymphoid tissue tumours [9]. One patient had DLBCL with features of classical Hodgkin lymphoma. Colon carcinoma and reactive hyperplastic lymphoid tissues, including lymph nodes and tonsils, were used as controls.
been reported in the past decade, suggesting a prevalence of P16 overexpression in human neoplasms [5–8]. Diffuse large B cell lymphoma (DLBCL) is a highly heterogeneous lymphoid malignancy comprising several disease subentities [9]. Based on cell origin, gene expression profiling using cDNA microarrays has identified three distinct subgroups of DLBCL: (1) the germinal centre B-cell-like (GCB) subgroup, with features of GCB cells, (2) the activated B-cell-like and type 3 subgroups, with features of post-GCB cells [10, 11], and (3) primary mediastinal B cell lymphoma (PMBL), with features of thymic B cells [12]. It was claimed that it is possible to identify DLBCL subgroups similar to those of cDNA microarrays by immunohistochemical analysis using a panel of antibodies [13]. In immunohistochemical analysis, expression of CD10 and BCL-6 is used to define the GCB phenotype whereas expression of MUM-1 is related to the non-GCB phenotype [13]. The GCB phenotypic subgroup is known as a predictor of a favourable outcome, while the non-GCB phenotypic subgroup appears to predict an unfavourable outcome [10, 13]. GCB and non-GCB DLBCLs may develop through different oncogenic pathways [14] and express differential levels of cell cycle regulators [15]. Loss of P16 in DLBCL has been studied by different research groups [15, 16] and has been shown to associate with the nonGCB phenotype [15]. In order to further elucidate the role of p16 in lymphomagenesis, we investigated P16 overexpression in de novo DLBCL by immunohistochemistry in order to clarify the sites of presentation, immunophenotypic subgroups and proliferation profiles as denoted by the expression of Ki-67. FISH was employed to analyse genetic alterations of p16 and their associations with high levels of P16 protein in DLBCL.
Immunohistochemistry The Envision system was employed for the staining of CD10, BCL-6 and MUM-1. The avidin-biotin complex system was employed for the staining of P16 and Ki-67. Three-micrometer-thick sections were cut from paraffin-embedded tissue blocks, deparaffinised in xylene and dehydrated through a graded ethanol series. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide in methanol. Heat-induced antigen retrieval was achieved by pressure cooking. After cooling to room temperature, antigen-retrieved sections were incubated with primary antibodies. Reactions against CD10 (1:50; clone 56C6, Novocastra Laboratories, UK), BCL-6 (1: 10; clone PG-B6p, DakoCytomation, Denmark), MUM-1 (1:25; clone MUM1p, DakoCytomation) and Ki-67 (1: 50; clone MIB-1, DakoCytomation) were run for 2 h at room temperature; overnight incubation at 4 ° C was performed for P16 (1:200; clone 16P07, Neomarkers, USA). Chromogenic development of all staining was achieved using diaminobenzidine (DakoCytomation). Haematoxylin was used as nuclear counterstain. External positive controls, i.e. reactive lymphoid tissues for staining of CD10, BCL-6, MUM-1 and Ki-67, and colon carcinoma for staining of P16 were included in every run of the assay. All stained slides were evaluated by a pathologist (S.C.P.) blinded to the clinicopathological data and FISH results. The percentages of tumour cells showing protein immunoreactivity were determined visually. CD10 was considered to be expressed when 630% of tumour cells showed membrane staining. BCL-6 and MUM-1 were considered to be expressed when 630% of tumour cells showed nuclear staining. The Ki-67 indices (percentages of cells expressing Ki-67) were stratified into five groups based on the extent of protein immunoreactivity: (1) !10% of positive tumour cells, (2) 10–25% of positive tumour cells, (3) 26–50% of positive tumour cells, (4) 51–75% of positive tumour cells, and (5) 175% of positive tumour cells. Staining was regarded as inconclusive if immunoreactivity was not observed in the internal positive controls, i.e. fibroblasts and neutrophils for CD10, and reactive lymphocytes for BCL-6, MUM-1 and Ki-67. The algorithm proposed by Hans et al. [13], based on CD10, BCL-6 and MUM-1 expression, was employed to stratify all the DLBCL cases into GCB and non-GCB phenotypic subgroups (fig. 1). P16 expression was separately classified into three categories: (1) simultaneous nuclear-cytoplasmic staining, (2) purely cytoplasmic staining, and (3) purely nuclear staining. Both the intensity and extent of staining were considered during the evaluation. Staining intensity was scored as follows: 1+ as weak, 2+ as moderate, and 3+ as strong. The extent of staining was graded as the percentage of positively stained tumour cells. Cut-off points for increased P16 expression were derived from seven reactive hyper-
p16 in Diffuse Large B Cell Lymphoma
Pathobiology 2010;77:96–105
Fig. 1. Stratification of DLBCL into GCB and non-GCB pheno-
typic subgroups based on the expression of three immunomarkers.
97
Table 1. Cut-off values for p16 alterations derived from means and standard deviations of normal cells with altered signals
Losses of p16
Mean 8 SD Cut-off values Specificity, % Sensitivity, %
monosomy 9
hemizygous deletion
homozygous deletion
polysomies 9
38.8488.67 56.19 [mean + 2 SD] 100 2
8.8582.74 17.07 [mean + 3 SD] 100 42
0.8681.02 4.96 [mean + 4 SD] 100 38
0.0780.18 0.61 [mean + 3 SD] 100 30
plastic lymphoid tissues. P16 overexpression was defined as 65% of tumour cells showing simultaneous nuclear-cytoplasmic staining, and/or 175% of tumour cells showing purely cytoplasmic staining, and/or 625% of tumour cells showing P16 staining with 3+ intensity score. Staining of normal lymphocytes and endothelial cells was used as internal positive controls. Fluorescence in situ Hybridization Interphase FISH analysis was performed using a commercially available dual-colour probe consisting of locus-specific identifier (SpectrumOrange) and centromere enumeration probe 9 (SpectrumGreen) (Abbott Molecular Inc., USA). Locus-specific identifier p16 spans approximately 190 kb and hybridizes to the 9p21 chromosomal region, containing the following genetic loci D9S1749, D9S1747, p16, p14, D9S1748, p15 and D9S1752. Centromere enumeration probe 9 is a chromosome 9 control probe designed to detect the ␣-satellite sequences 9p11–q11. The FISH assay was performed as previously described [17]. A reactive lymphoid tissue section was incorporated in every run of the assay as experimental control. All FISH specimens were evaluated using a fluorescence microscope equipped with appropriate filter cubes. Specimens were scanned using a 40! objective to identify the tumour areas. Digital images of the tumour areas were acquired using an integrating monochrome CCD camera and a 100! objective. At least 190–210 intact and non-overlapping nuclei were scored for the copy numbers of p16 (visualized as red signal) and chromosome 9 (visualized as green signal). Only nuclei with at least 1 green signal were enumerated. The scoring system was based on the percentage of nuclei showing different copy numbers of p16 and chromosome 9. An additional slide was prepared for specimens with less than 190 evaluated nuclei. Cases were considered noninformative if the supplemental slide failed to yield sufficient nuclei for analysis. Cases with poor signal intensity and/or excessive background autofluorescence were also regarded as non-informative. Cut-off values for p16 alterations were determined from 15 normal controls (reactive hyperplastic lymphoid tissues) based on the mean percentages of normal cells exhibiting altered signals and dispersions (as defined by standard deviations) of the incidence values (table 1). All established cut-off values were verified in terms of specificity and sensitivity using receiver operating characteristic curve analysis. Thresholds with 100% specificity were considered. Cases with altered signals exceeding the corresponding cut-off values were interpreted as alterations of p16.
98
Gains of p16
Pathobiology 2010;77:96–105
Statistical Analysis Pearson’s 2 and Fisher’s exact probability tests were used to examine the relationships between the different investigated parameters. The Mann-Whitney U test was employed to determine the differences in the frequencies of p16 alterations between normal controls and DLBCL. The results were considered statistically significant when p ! 0.05. SPSS for Windows, version 13 (SPSS Inc., USA) was used to perform all statistical analysis.
Results
Patient Data The DLBCL series included 38 (54.3%) male and 32 (45.7%) female patients. The mean age of these patients at the time of diagnosis was 52.4 8 16.5 years (median, 53.5 years; range, 3–86 years). 40.0% (28/70) of patients were aged 50 years or less. The studied subjects were comprised of three ethnic groups: Chinese (33/70, 47.1%), Malay (28/70, 40.0%) and Indian (9/70, 12.9%). The clinicopathological features of each patient are presented in table 2. Immunophenotypic Profiles Of all the DLBCL cases, 32.9% (23/70) exhibited positive CD10 expression, 62.9% (44/70) expressed BCL-6 and 71.4% (50/70) demonstrated positive MUM-1 expression. 38.6% (27/70) of the cases were assigned to the GCB subgroup, and 61.4% (43/70) to the non-GCB subgroup (fig. 1). Proliferation Profiles The cell proliferation marker Ki-67 was expressed in all (100%) DLBCL cases; positivity ranged from 10 to 175%. No cases had Ki-67 indices !10%. Ki-67 indices of 10–25, 26–50 and 51–75% were noted in 11.4% (8/70), 21.4% (15/70) and 22.9% (16/70) of the cases, respectively.
Lee /Kim /Abdullah /Peh
Color version available online
Fig. 2. Overexpression and alterations of p16 gene in DLBCL. a Overexpression of P16 protein is shown in both nucleus and cytoplasm of the tumour cells. !25. b Hemizygous deletion of p16 gene is in the tumour cells (arrowheads). !100. c Homozygous deletion of p16 gene shown in the tumour cells. !100. d Tumour cells revealed gains of p16 gene with trisomy and tetrasomy of chromosome 9 (arrows). !100.
The majority (31/70; 44.3%) of the cases showed strong Ki-67 expression with indices higher than 75%.
Alterations of p16 FISH signals were interpretable in 67.1% (47/70) of the DLBCL cases. The Mann-Whitney U test demonstrated that the frequencies of monosomy 9 (p = 0.019), hemizygous p16 deletion (p = 0.002), homozygous p16 deletion (p ! 0.001) and polysomies 9 (p = 0.032) were significantly different between DLBCL cases and normal controls.
There was no significant difference between tumour cases and normal controls in the frequencies of p16 gene amplification (p = 0.771). Monosomy 9 and hemizygous and homozygous deletions were referred to as losses of p16. Gains of p16 in the present DLBCL series were observed as polysomies 9. Of the 47 interpretable hybridizations, hemizygous and/or homozygous deletions were evident in 27 (57.4%) cases. The concurrent occurrence of hemizygous and homozygous deletions in individual cases was statistically significant (p = 0.043). Simultaneous hemizygous and homozygous deletions were present in 11 of 27 (40.7%) cases, purely hemizygous p16 deletion was shown in 9 of 27 (33.3%) cases, and purely homozygous deletion was shown in 7 of 27 (25.9%) cases. None of the examined specimens showed monosomy 9. Gains of p16 due to polysomies 9 were noted in 27.7% (13/47) of the cases. Representative examples of DLBCL harbouring cytogenetic alterations are illustrated in figure 2b–d. p16 deletions were significantly associated with extranodal presentation (p = 0.047). No statistically significant associations or differences were observed in different subgroups of DLBCL relative to p16 deletions (p = 0.137) and p16 gains (p = 0.726). P16 protein overexpression was not significantly associated with gains of p16 gene (p = 0.621), but was inversely related to p16 gene deletions (p = 0.019) (table 4).
p16 in Diffuse Large B Cell Lymphoma
Pathobiology 2010;77:96–105
Expression of P16 P16 expression was interpretable in all (100%) DLBCL cases. Overall, 45.7% (32/70) of the cases showed P16 overexpression. High levels of P16 were observed predominantly as simultaneous nuclear-cytoplasmic staining (29/32; 90.6%) as shown in figure 2a or purely cytoplasmic staining (3/32; 9.4%). P16 overexpression and CD10 positivity (p = 0.022) were significantly associated. Increased P16 expression was significantly associated with the GCB subgroup (p = 0.022). An inverse correlation was found between P16 overexpression and a high proliferative index (Ki-67 index 175%) of tumour cells (p = 0.020). No statistically significant correlations were found between P16 overexpression and sites of presentation (p = 0.826) and expression of BCL-6 (p = 0.152) and MUM-1 (p = 0.129) (table 3).
99
Table 2. Patient data, sites of presentation, phenotypic subgroups, P16 overexpression and genetic alterations in
DLBCL
100
Case No.
Age years
Gender
Ethnic group
Site of presentation
Phenotypic subgroup
P16 overexpression
p16 alteration
1 3 4 7 8 10 11 12 14 15 18 19 23 26 27 29 31 32 34 36 37 38 40 41 43 46 47 51 53 54 56 57 58 60 2 6 9 28 33 52 65 68 70 5 13 25 44 48 50 55 17 64 66
54 38 61 62 32 32 12 25 53 52 65 58 61 38 60 64 49 34 69 74 54 43 63 46 43 29 63 84 49 3 39 74 51 42 63 58 31 57 63 52 60 56 70 53 51 81 40 59 64 26 86 65 45
M M F M F M M M F M M F F F F F M M M M M F M F F M F F M F M M M M F F M F F M M F F F F M F M M M F F F
C C M M M C C C M C I C C C I C M C C C C M C M M M M I C C C I C M C C M C M C M M M I C C C M I C M M C
LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN LN tonsil tonsil tonsil tonsil tonsil tonsil tonsil tonsil tonsil GI tract GI tract GI tract GI tract GI tract GI tract GI tract nasal region nasal region nasal region
GCB non-GCB non-GCB non-GCB non-GCB GCB GCB GCB non-GCB GCB GCB non-GCB non-GCB non-GCB GCB GCB non-GCB GCB non-GCB non-GCB non-GCB non-GCB GCB GCB non-GCB non-GCB non-GCB GCB GCB GCB GCB non-GCB GCB non-GCB non-GCB GCB non-GCB non-GCB non-GCB GCB non-GCB non-GCB non-GCB GCB non-GCB non-GCB non-GCB non-GCB non-GCB GCB non-GCB GCB non-GCB
+ + – + – – + – – + + – + – – + – + – + – – – + – + – + + + – – + – + + + – + – – – + + + – + – – + – – –
NI G NI NI NA D NI G G NA NA G NA D D D NA D G D D D D/G G D NI NI NI G NA NA D NI NI NI NA NI D NA NI D D/G D NA NI D G NI NI D D NI NI
Pathobiology 2010;77:96–105
Lee /Kim /Abdullah /Peh
Table 2 (continued)
Case No.
Age years
Gender
Ethnic group
Site of presentation
Phenotypic subgroup
P16 overexpression
p16 alteration
67 69 24 30 20 42 16 21 22 35 39 45 49 59 61 62 63
58 48 40 77 56 83 61 74 29 36 54 36 71 37 51 49 49
F M F M M M M M M M M F M F F M F
I M M M C C M M I I C M M C M M C
nasal region nasal region bone bone skin skin testis thyroid lung mediastinum brain oropharynx spinal-cauda equina submandibular gland neck region soft tissue breast
GCB GCB non-GCB non-GCB non-GCB non-GCB non-GCB GCB non-GCB non-GCB GCB GCB non-GCB non-GCB non-GCB GCB non-GCB
– + + – + – – – + + – + – – – + –
D D D/G D NI NI D D NI NI NA NI D D D/G G NI
M = Male; F = female; C = Chinese; M = Malay; I = Indian; LN = lymph node; GI = gastrointestinal; GCB = germinal centre B-cell-like; G = gain; D = deletion; NA = no alteration; NI = non-informative.
Table 3. P16 overexpression in DLBCL in relation to sites of presentation, expression of CD10, BCL-6 and MUM-1, phenotypic subgroups and proliferation profiles (Ki-67 indices)
Overexpression of P16
Site of presentation, n Nodal Extranodal CD10, n Positive (≥30%) Negative (
